(May 12, 2023 at 2:33 am)emjay Wrote: I'm basically kicking myself for not trying it sooner

You know, if I were still writing academic papers I think I would use this. 

The hardest part in such a paper is the first draft, and then it's just a question of clarification and refinement. If the AI did the first draft for me, I could go through adding and subtracting until it was mine. 

Although there is the question of where the AI gets its information from. Somehow it searches around and gets information about Swedenborg, for example -- I don't think it has such information ready in its memory. So if I was writing a paper with the ambition of adding original findings to the world, the first draft wouldn't be enough. You could use it for things like background, though. Like the already-accepted points which you're holding up as evidence for your new proposition.

(May 12, 2023 at 3:27 am)Belacqua Wrote:

(May 12, 2023 at 2:33 am)emjay Wrote: I'm basically kicking myself for not trying it sooner

You know, if I were still writing academic papers I think I would use this. 

The hardest part in such a paper is the first draft, and then it's just a question of clarification and refinement. If the AI did the first draft for me, I could go through adding and subtracting until it was mine. 

Although there is the question of where the AI gets its information from. Somehow it searches around and gets information about Swedenborg, for example -- I don't think it has such information ready in its memory. So if I was writing a paper with the ambition of adding original findings to the world, the first draft wouldn't be enough. You could use it for things like background, though. Like the already-accepted points which you're holding up as evidence for your new proposition.

Personally I'd be very cautious about publishing anything produced with it, given the concerns raised in this video:





Basically, since for instance you can't know how much of its output is synthesised vs quoted verbatim, you'd always be at risk of inadvertantly breaching someone else's copyright if it happened to quote a part of their work, and thus according to these terms and conditions, being held liable for that if they decided to sue either you or openai (as I understand it). So as much as I want to share its creative output, I find myself reluctant to on account of this, instead preferring to err on the side of caution.

Also, here's the video I mentioned earlier about how it works:





So from that, it basically appears to me to essentially be a large-scale predictive text model, whose output can be influenced by context, not just from within the latest prompt, but also from the entire conversation leading up to it. So in that sense it appears it can produce novel outputs, as constrained by the specific context, such as in Neo's case, the constraints of Swedenborg and Terriers... the combination of those two would be unique enough that the output would also in one sense be 'novel' ie how the neural network 'settles' to accomodate those particular constraints, such that a unique combination of constraints will likely settle into a similarly unique configuration of the network, but at the same time I think brewer has hit upon an interesting point in his question to Neo, though his point may be different from mine, but basically the question of how determined that output is. Ie whether the same given set of constraints will always produce the same output, or whether there's any extra variation added to that process. The fact that it ranks outputs is one sort of variation I suppose, but I wonder if there are others... ie for instance if there's any RNG involved.

Last week I had a relatively simple math problem I didn't know the formula for so I asked ChatGTP. The question I asked was, "How much is the first year of a 5 year contract with a total value of 100000000 and a 2.5% annual escalation rate worth?"

It gave me an obviously wrong answer. I told it that was wrong. It told me I was was right the answer was wrong. Then it gave me the same wrong answer again. I tried to rephrase the question several times but kept getting bad answers.  

After a few minutes of arguing with a computer, I decided to figure it out for myself which I did using the -PMT formula in Excel. I went back to ChatGTP and told it the correct answer but not how I got there. It finally came up with a formula that returned the correct answer. Then I asked it for the Excel formula to do the calculation and it proceeded down another rabbit hole of wrong answers until I once again had to tell it the correct formula.

I went back to ChatGTP this morning before starting this post and asked the same question again. Apparently it can't remember shit because I got another wrong answer. This morning's answer was $17,805,580.20. The correct answer is $19,024,686.09. My conclusion from all this is a bunch of kids are going to be sorely disappointed when they get bad grades on their math home work.

Wasn't enough meanness in any of the ridicule posts, but as I understand it, that's intentional, as otherwise it would quickly devolve into a hatespeechbot.

(May 12, 2023 at 8:18 am)The Grand Nudger Wrote: Wasn't enough meanness in any of the ridicule posts, but as I understand it, that's intentional, as otherwise it would quickly devolve into a hatespeechbot.

Yes, on talking to it about its inner workings, it confirms that the largely static dataset it works from, was 'curated' and that curation process... of preparing the data for inclusion in the model... includes filtering out inappropriate/offensive content, including profanity, as well as sensitive information such as personally identifying information etc. So it's basically filtered at source and that's why it's so friendly

It doesn't understand relationships either, it's response from the point of view of atheists or the boston terrier society were effectively identical. Basically a fancy form letter generator. As I understand it, this is why it's not actually a huge issue for educators. Too easily identified either by eye or by any number of other programs. One kid using it might slip passed. Two kids using it reduce each others chances. More makes it entirely obvious.

(May 12, 2023 at 9:05 am)The Grand Nudger Wrote: It doesn't understand relationships either, it's response from the point of view of atheists or the boston terrier society were effectively identical.  Basically a fancy form letter generator.  As I understand it, this is why it's not actually a huge issue for educators.  Too easily identified either by eye or by any number of other programs.  One kid using it might slip passed.  Two kids using it reduce each others chances.  More makes it entirely obvious.

You seem somewhat nonplussed by it lol? Do you use it, or did you use it, maybe the novelty has worn off for you? I still think it's amazing and can imagine myself getting really addicted to using this, to clarify my thinking and all sorts. For instance I'm not great at let's say the management of learning or just how to get started etc... ie the bigger picture... but I think things like this could really help.

(May 12, 2023 at 8:57 am)emjay Wrote: Yes, on talking to it about its inner workings, it confirms that the largely static dataset it works from, was 'curated' and that curation process... of preparing the data for inclusion in the model... includes filtering out inappropriate/offensive content, including profanity, as well as sensitive information such as personally identifying information etc. So it's basically filtered at source and that's why it's so friendly

I suspect that people write or say offensive content, or profanity, or other nasty attempts to make other people feel bad, because they feel some strong desire. Which the AI of course lacks. 

Since attacking others in writing changes no one and improves nothing, it's only done to make the attacker feel good -- to satisfy some desire they have. The boffins who program the AI at a distance from any sort of debate have the benefit of not feeling that kind of desire while they're at work. 

And unlike AI text, art, too, is based on desire, though of a different kind:

"...aesthetic bliss, that is a sense of being somehow, somewhere, connected with other states of being where art (curiosity, tenderness, kindness, ecstasy) is the norm." -- Nabokov

(May 12, 2023 at 9:26 am)Belacqua Wrote:

(May 12, 2023 at 8:57 am)emjay Wrote: Yes, on talking to it about its inner workings, it confirms that the largely static dataset it works from, was 'curated' and that curation process... of preparing the data for inclusion in the model... includes filtering out inappropriate/offensive content, including profanity, as well as sensitive information such as personally identifying information etc. So it's basically filtered at source and that's why it's so friendly

I suspect that people write or say offensive content, or profanity, or other nasty attempts to make other people feel bad, because they feel some strong desire. Which the AI of course lacks. 

Since attacking others in writing changes no one and improves nothing, it's only done to make the attacker feel good -- to satisfy some desire they have. The boffins who program the AI at a distance from any sort of debate have the benefit of not feeling that kind of desire while they're at work. 

And unlike AI text, art, too, is based on desire, though of a different kind:

"...aesthetic bliss, that is a sense of being somehow, somewhere, connected with other states of being where art (curiosity, tenderness, kindness, ecstasy) is the norm." -- Nabokov

So what's your point? What argument are you trying to make here?

Or maybe human beings use expletives to be more expressive and place emphasis, which the chatbot also doesn't know about. I wonder, if it tried, would it sound like a russian trying to curse in americaneese?

@emjay

Yeah, I mean, chatbots are cool, but they're just chatbots. The stuff they tell you is whatever stuff is trending across the internet. Often enough, the stuff trending across the internet isn't true. They'll make up entire cities, and without extremely limiting guardrails they breathlessly expound on the virtues of white supremacy when you ask them how to make an omelette or grill kabob. When they're fully deployed in the wider world they aren't going to say anything that isn't part of some online retailers ad copy....so I wouldn't count on them to clarify any thought I had unless I was looking to buy a timeshare in the adriatic.